I>aphnia galeata mendotae as a cryptic species complex with interspecific hybrids

Transcription

1 658 Notes scattering layers and its significance to surface swarms. J. Plankton Res. 5: SIMARD,~., R. DELADURANTAYE,AND J. THERRIAULT Aggregation of euphausiids along a coastal shelf in an upwelling environment. Mar. Ecol. Prog. Ser. 32: WALDRON, D. E Trophic biology of the silver hake (Merluccius bilinearis) population on the Scotian Shelf. Ph.D. thesis, Dalhousie Univ. WIDDER, E. A., AND OTHERS Bioluminescence in the Monterey Submarine Canyon: Image anal- ysis of video recordings from a midwater submersible. Mar. Biol. 100: WIEBE, P. H., C. 11. GREENE, T. K. STANTON, AND J. BURCZYNSKI Sound scattering by live zooplankton and micronekton: Empirical studies with a dual-beam acoustical system. J. Acoust. Sot. Am. 88: Submitted: 27 December I990 Accepted: 23 September 1991 Revised: 3 February 1992 Limnol. Oceanogr., 37(3), , by the American Society of Limnology and Oceanography, Inc. I>aphnia galeata mendotae as a cryptic species complex with interspecific hybrids Abstract- Daphnia galeata mendotae has been recognized as possessing the greatest morphological variation among North American Daphnia species. Prior work has suggested that this variation is a consequence of phenotypic plasticity. Using allozyme markers, we tested this hypothesis by examining the association between genotype and morphology in 22 populations of D. galeata from lakes in Indiana. Analyses indicated a genetic basis for the phenotypic variability, as two different species and their F, hybrids were detected. The parent species were D. galeata mendotae sensu strict0 and a helmeted form of Daphnia rosea. F, hybrids were detected in 17 of 22 lakes and were dominant in 10. Backcross and F, individuals were rare, however, suggesting that introgression is limited by the existence of reproductive isolating mechanisms. Despite its importance in freshwater habitats, the taxonomy of the crustacean genus Daphnia remains confused. Brooks (19 57a) proposed that difficulties in species assignments resulted from extensive variation created by a combination of phenotypic plasticity, the coexistence of morphologically similar species, and interspecific hybridization. Genetic studies have now esta.blished that these three factors have -- Acknowledgments We thank W. R. DeMott for use of the Crooked Lake Field Station; M. Beaton, R. De Melo, T. Finston, P. Gajda, and T. Van Raay for aid in the collection of samples; and P. Forde for the Daphnia illustrations. This research was supported by a Natural Sciences and Engineering Research Council grant to P.D.N.H. interacted to create variation in both Australian (Hebert 1985) and European Daphnia species groups (Wolf and Mort 1986; Hebert et al ). In North America, prior work has shown that Daphnia galeata Sars, 1864 mendotae Birge, comb. nov. Brooks, 1957 possesses a striking amount of phenotypic variation (Brooks 1957b; Jacobs 196 1; Mort 1989). Indeed, YBrandlova et al. (1972) suggested that D. galeata mendotae might represent a species group because its morphological variation was the greatest among North American Daphnia species. But stud- ies of environmentally induced morphological changes (Jacobs 196 1) and of allozyme variation (Stirling and McQueen 1987; Mort 1989) have failed to provide evidence for the species group proposal. In fact, all of these investigations have concluded that genetic sources of morphological variation in D. galeata mendotae are minor and that phenotypic plasticity is great. However, such conclusions seem premature. The genetic studies were concerned mainly with allozyme-morphology associations within a single population and were therefore inadequate for assessing associations at the species level. Because multilake studies (e.g. Brandlova et al. 1972) have shown that phenotypic variation is much greater among lakes than within lakes, prior allozyme studies have examined only a small fraction of the total morphological variation

2 Notes 659 in the species. Furthermore, the probability of detecting an association between genotype and phenotype increases with the number of allozyme marker loci used. Prior studies have used only two to four marker loci in their search for genotype-phenotype covariation. Our purpose was to provide a more comprehensive evaluation of the sources of phenotypic variation in D. galeata mendotae. To do this, we examined the association between morphology and allozyme genotypes at 16 different loci among 22 populations of D. galeata mendotae. Samples of Daphnia were obtained from 24 glacial lakes in three watersheds (Tippecanoe, Wabash, and St. Joseph) and two drainage basins (Lake Michigan and Ohio River) in northeast Indiana. Three morphs are prevalent in this region (Fig. 1). The first possesses an apically pointed helmet and a concave ventral head margin, while the second, Woltereck s (1932) Daphnia longispina elongata indianae, has a rounded helmet and a convex ventral head margin. Brooks ( 1957 b) synonymized these two morphs with D. galeata mendotae. The third morph is intermediate in these characters and also agrees with the definition of D. galeata mendotae. The morphological relationships of males and juvenile females follow the same pattern as the adult females. Morph 1 males and juvenile females always have an apical spike, whereas morph 2 males and juvenile females are always apically rounded. Morph 3 has an intermediate phenotype, with males always, and juvenile females frequently, possessing an anterodorsal denticle. A more posterior and roseate neck-tooth develops in some males of morph 2. Another closely related species, Daphnia rosea Sars, 1862, has been identified from ponds in Indiana (pers. obs.) and is discriminated from D. galeata by its lack of a helmet (Brooks 1957b). In all, 22 populations of D. galeata and 2 of D. rosea were analyzed. Samples were collected by duplicate vertical hauls of a 30- cm-diameter plankton net (200-pm mesh) near the point of maximum depth in the lake. D. galeata was sorted from other zooplankton and within 5 h of collection either frozen in liquid nitrogen, or preserved in 95% ethanol. Before electrophoresis, individuals were characterized as one of the three morphs. Cellulose acetate gel electrophoresis was conducted according to the methods of Hebert and Beaton (1989). The following 12 enzyme systems and 16 loci were scored: aldehyde oxidase (AO), arginine phosphokinase (Apk), dipeptidase (Pep-A, Pep-D), fumarate dehydrogenase (Fum), glutamate-oxaloacetate transaminase (Gotm, Got-s), glyceraldehyde-3-phosphate dehydrogenase (G3pdh), isocitrate dehydro- genase (Idh- I, Idh-2), lactate dehydrogenase (Ldh), malate dehydrogenase (Mdh- m, Mdh-s), malic enzyme (Me), phosphoglucomutase (Pgm), and phosphoglucose isomerase (Pgi). Substrates for Pep-A and Pep-D were L-leucylglycine and L-phenylalany1 proline respectively. Relative allozyme mobilities were calculated as a percentage of the mobility of the most common D. rosea allele. A locus was considered polymorphic when the most common allele had a frequency of ~95%. Data from May 1990 were used for genetic analysis of most populations, but data from September 1989 and May 1990 were pooled for five populations of morph 2 (following a G-test of independence) because individual sample sizes for this morph were small in May. Three populations of morph 3, for which only seven loci were available, were excluded from the hetero- zygosity calculations. Seven (Apk, G3pdh, Idh-1, Idh-2, Mdhm, Mdh-s, Me) of the 16 loci were invariant among all individuals, but the other nine were useful in discriminating between morphs 1 and 2 (Table 1). Fixed allelic differences between these morphs occurred at two loci (AO, Got-m), and large gene frequency differences occurred at four additional loci (Got-s, Pep-A, Pep-D, Pgm). Three other loci were partially diagnostic: morph 1 possessed alleles (Fuma, and Pgib) that were not detected in morph 2, and morph 2 possessed an allele (Ldh ) that was not detected in morph 1. The allozyme survey suggested that individuals of morph 2 were closely related to D. rosea as they had identical allelic arrays at all loci except Ldh, where morph 2 populations possessed a unique allele (Ldha).

3 Morph 1 Morph 2 Morph 3 I / / a I P h

4 Notes 661 Table 1. Mean allele frequencies (AF) for morph 1, morph 2, and Daphnia rosea, and mean genotype frequencies (GF) for morph 3. Frequencies represent means of eight lakes for morph 1, nine for morph 2, and two for D. rosea. Locus AO Fum Got-m Got-s Ldh Pep-A Pep-D Pgi pm Alleles (relative mobility) a (100) b (142) a (88) b (100) a (100) b (127) a (100) b (114) a W3) b (100) a (100) b (121) a (81) 6 (100) b (100) c (107) a (89) b (100) c (109) Morph 3 Morph 1 Morph 2 D. rosea Obsewed No. of AF N AF N AF N genotypes GF N lakes oo oo 325 ab ab oo 1.oo bb - 1.oo ab oo aa ab ab bb oo aa ab ab oo 1.oo bb ab bb bc ab z bc 1.ooo ooo Individuals of morph 3 were invariably heterozygous at AO and Got-m (Table 1). These heterozygotes consisted of one allele typical of morph 1 and one typical of morph 2. At the remaining polymorphic loci, all but one allele (Pgi ) in morph 3 individuals was detected in either morph 1 or in morph 2. This allele was, however, detected in morph 1 individuals from Michigan (D. Taylor unpubl.). The observed genotypes and their frequencies at all 16 loci were consistent with the hypothesis that morph 3 was an F1 hybrid between the other two morphs. Morph 2 was polymorphic at a lower proportion (X = lo.o%, range O-12.5%) of the 16 loci than morph 1 (2 = 2 1.9%, range %). Individuals of morph 3, however, showed extremely high levels of polymorphism (jz = 38.4, range %). Differences in mean heterozygosity (H) over 16 loci mirrored those in polymorphism (Fig. 2). Populations of morph 3 (H = 0.337) had over 3 times greater heterozygosity than those of morph 1 (H = 0.10 l), and 14 times greater heterozygosity than those of morph 2 (H = 0.024). These differences provided a pattern of nonoverlapping heterozygosity estimates for the three morphs. In May 1990, F1 hybrids (i.e. morph 3) were detected in 17 of 22 lakes and were the dominant morphs in 10 (Fig. 3). Four lakes contained only hybrids, and on average, 44.5% of individuals collected in the 22 lakes were hybrids. Hybrids coexisted with either one or the other of the putative parent morphs in 12 lakes and with only a backcross/f2 genotype in one lake. In three lakes (Blue, Kuhn, and Loon), hybrids were represented by single unique clones, while in seven others (Adams, Bear, Crooked, Lime, Lower Long, Wawasee, and Webster) hybrid populations consisted of two to five t Fig. 1. Morphs of Daphnia galeata mendotae from the Indiana lake district. Mature females: a-center Lake, 3 October 1990; b-smalley Lake, 2 October 1990; c-tippecanoe Lake, 3 October Immature females: d-silver Lake, 8 May 1990; e-bear Lake, 6 May 1990; f-bear Lake, 6 May Males: g-center Lake, 3 October 1990; h-old Lake, 3 October 1990; i-blue Lake, 6 May (Scale bars- 1 mm.)

5 662 Notes 6 5 n Morph 1 Cl Morph 2 EI Morph 3 g 4.I z 3 m 0 3 e % 1 g 02 : 0.00 Mean Heterozygosity Fig. 2. Mean heterozygosity at 16 loci for three morphs of Daphnia galeata in 22 Indiana lakes. clones each. The other seven hybrid populations were either clonally diverse (> 5 clones) or sample sizes were too small to adequately assess clonal diversity (Little Crooked and Waubee). Putative backcrosses or F,s, which were originally grouped with morph 3, were detected in four lakes. These individuals were identified by their possession of both recombinant hybrid genotypes and homozygous parental genotypes at diagnostic loci. At three sites, such genotypes included < 10% of the population samples, but in one lake (Waubee), a single backcross clone constituted 9 1% of the complex. Among the remaining lakes, morph 1 dominated six lakes, while morph 2 dominated five. Hybrids as well as both parent morphs were detected in each of the three watershed areas, but the morphs differed with respect to the size of lake in which they occurred. Morph 1 occurred in lakes of larger surface area -[mean log(surface area in ha) = than the lakes in which morph 2 occurred [mean log(surface area in ha) = , t = 3.0 1, df = 15, P = Although the lake size distributions overlapped for the parent morphs, we never detected them coexisting. Instead, the lake size classes that overlapped between the parent morphs were dominated by the hybrid morph. The mean area of lakes occupied by hybrids [log(ha) = was intermediate to those of the parent morphs, but differed significantly only from morph 2 (t = 2.07, df = 24, P = 0.05). D. rosea was detected in the two smallest lakes (Hammond and Allen) of the sample area. Even though lake depth (m) was correlated with surface area in the study area (r = 0.47, P = O.OOS), no significant differences were found among depths of lakes in which different morphs occurred. Allozyme analysis indicated that hybrids were less common in fall than in spring, being detected in only 9 of 15 lakes and comprising 18O/o of all individuals. The reduction in the relative abundance of hybrids was due to an increase in the abundance of morph 2 in small-medium-sized lakes. In larger lakes, however, hybrids actually in- creased in abundance relative to morph 1. This study indicates that the phenotypic

6 Notes 663 Fig. 3. Relative abundance of Daphnia guleata morphs (based on allozyme analyses) in May 1990 for 22 Indiana lakes. Sample sizes range from 35 to 168 with a mean of 79. Lake codes: AD-Adams; X-Big Crooked; BG-Big; BL-Blue; BR-Bear; CN-Center; CR-Crane; HG-High; JM-James; KN-Kuhn; LC-Little Crooked; LL- Lower Long; LM- Lime; LN- Loon; OL- Old; SL- Silver; SM- Smalley; TP- Tippecanoe; WB - Waubee; WL- Waldron; WW - Wawasee; WS - Webster. variability present among members of the D. galeata complex from Indiana has a strong genetic basis. It is clear that two species and their interspecific hybrids occur in these lakes. Our results suggest that only morph 1 individuals should be assigned to D. galeata mendotae. This was the only morph that possessed a tall, pointed helmet-a character that, according to Brooks ( 1957 b), separates D. galeata mendotae from closely related species. The presence of fixed allozyme differences and a moderate genetic divergence (Nei s unbiased genetic distance = for 16 loci) between D. galeata mendotae (i.e. morph 1) and morph 2 suggests that they are separate, closely related species. On the other hand, the near identical allozyme arrays and morphological similarities between morph 2 and D. rosea sug- gest that these two taxa are conspecific. Morph 2 shares with D. rosea the following characters: a slender tailspine, neckteeth, pinkish bodies, and a habitat preference for small lakes. Based on these similarities, we propose that morph 2 is a helmeted form of D. rosea. Brooks (1957b) grouped morph 2 with D. galeata mendotae because of its morphological and ecological deviance from typical D. rosea. He described North American D. rosea as a western species that was most common in ponds and ordinarily lacked a cephalic crest. Morph 2 s possession of a cephalic crest and its occupancy of lakes in the eastern half of North America provided Brooks (19573) with the justification for its fusion with D. galeata mendotae. However, it is now well known that D. rosea occurs throughout most of North America (Brandlova et al. 1972). Moreover, Brooks (1957b) acknowledged that robust forms of D. rosea could develop crests (see the robusta and

7 664 Notes dinotocephala synonyms of Kiser). More work is required to determine if morph 2 is a genetically divergent form of D. rosea or simply a phenotypic response to the unique conditions of Indiana marl lakes. These lakes are warmer (being at the southern extent of glaciation), more turbid (due to colloidal CaCO,), and contain higher concentrations of calcium than other North American glacial lakes (Wetzel 1966). Genetic evidence for interspecific hybridization between North American Daphnia has been obtained in only one case (Daphnia pulex X pulicaria, Hebert et al. 1989a). Morph 3, being intermediate with respect to D. galeata and D. rosea in genetic structure, morphology, and habitat occupancy, provides a second clear example of Daphnia hybridization in North America. Hybrids between members of the D. galeata complex and other sympatric Daphnia species (ambigua, longiremis, pulicaria, and retrocur- VCZ), could readily be recognized by fixed allozyme markers that exist for each of these species. However, as no heterozygotes were detected at these marker loci, we conclude that there is no hybridization between any of these species and members of the D. galeata group in Indiana (D. Taylor unpubl.). The abundance of Daphnia hybrids from Indiana (at 44.5% in the spring) is the greatest yet recorded for lakes. F1 hybrids constituted 2 1% of the D. galeata complex in Germany (Wolf and Mort 1986) and 1.3% of the complex in Czechoslovakia (Hebert et al ). Both of these studies did, however, find water bodies in which F1 hybrids were temporarily dominant. Previous work has suggested that the hybrids present in many European lakes were the result of multiple hybridization events, as hybrid swarms were genotypically diverse (Hebert et al ). Moreover, Wolf and Carvalho (1989) found that F, hybrids hatched from ephippia collected from lake sediments. By contrast, the low number of multilocus genotypes detected within most Indiana hy- brid populations suggests that high densities are often achieved by parthenogenetic replication of single hybrids. Given the parental variation, the evidence for asexual amplification of hybrids is particularly compelling in lakes where only one geno- type was detected. Selection for certain hybrid clones may also have reduced diversity. Yet, as uniclonal samples were obtained at the probable peak of ephippial hatching (Wolf and Carvalho 1989), it is unlikely that selection accounts for the low diversity. In contrast to European lakes, in which cooccurrence of the parent species is common, the low frequency of hybridization events in Indiana lakes is almost certainly a consequence 0.f the infrequent coexistence of both parent taxa. A more challenging problem than the origins of hybrids is the maintenance of the entire hybrid complex. If hybrid forms were more fit than either parent taxa, then replacement of the parent species might occur on a regional scale. Such replacement did not occur in Indiana because of differing habitat preferences for the parent taxa and their hybrids. 13. rosea was dominant in small lakes, hybrids were dominant in medium-sized lakes, and D. galeata was dominant in large lakes. These results support Lieder s (1983) suggestion that Daphnia hybrids have optimum habitats intermediate between those of the parents. Their differing habitat preferences provide an ecological basis for persistence of the parent taxa, but hybridization might, if coupled with introgression, cause convergence of the parent forms. Although the present study detected both F, males and ephippial F1 females, the number and abundance of backcross genotypes were low. There was, however, evidence for gene flow from D. rosea to D. galeata because uncommon alleles at three loci (Pep-A, Pep-D, Got-sa) in D. galeata were the sole alleles present in D. rosea. Furthermore, the higher level of heterozygosity in D. galeata compared to D. rosea may, in part, be a product of this introgression. It seems unlikely that hybridization in the D. galeata complex occurs only in Indiana. Several studies have figured animals from other parts of North America which bear a striking resemblance to the F1 hybrids examined in this study (e.g. Woltereck 1932). These animals have often been interpreted as stages in a cyclomorphic succession. The present study reveals that hybrids are particularly cryptic when they coexist with D.

8 galeata. In such lakes, the temporal succession of hybrids and D. galeata creates a continuum of head shapes that is readily interpreted as cyclomorphosis. Clearly, the prevalence and geographic distribution of hybrids needs to be more thoroughly investigated. Department of Zoology University of Guelph Guelph, Ontario N 1 G 2W 1 References Derek J. Taylor Paul D. N. Hebert BRANDLOVA, J., Z. BRANDL, AND C. H. FERNANDO The Cladocera of Ontario with remarks on some species and distribution. Can. J. Zool. 50: BROOKS, J. L. 1957a. The species problem in freshwater animals, p In E. Mayr [ed.], The species problem. Am. Assoc. Adv. Sci b. The systematics of North American Daphnia. Mem. Conn. Acad. Arts Sci. 13: 180 p. HEBERT, P. D. N Interspecific hybridization between cyclic parthenogens. Evolution 39: AND M. J. BEATON Methodologies for aliozyme analysis using cellulose acetate electrophoresis. Helena Laboratories. - -, S.S.SCHWARTZ, ANDD.J.STANTON. ld89a. Polyphyletic origins of asexuality in Notes 665 Daphnia pulex. 1. Breeding system variation and levels of clonal diversity. Evolution 43: 1004-l , S. S. SCHWARTZ, AND J. HRBACEK. 1989b. Patterns of genotypic diversity in Czechoslovakian Daphnia. Heredity 62: JACOBS, J Cyclomorphosis in Daphnia galeata mendutae, a case of environmentally controlled allometry. Arch. Hydrobiol. 58: LIEDER, U Introgression as a factor in the evolution ofpolytypical plankton Cladocera. Int. Rev. Gesamten Hydrobiol. 68: MORT, M. A Cyclomorphosis in Daphnia galeata mendotae: Variation and stability in phenotypic cycles. Hydrobiologia 171: 159-l 70. STIRLING, G., AND D. J. MCQUEEN The cyclomorphic response of Daphnia galeata mendotae: Polymorphism or phenotypic plasticity. J. Plankton Res. 9: 1093-l 112. WETZEL, R. G Productivity and nutrient relationships in marl lakes of northern Indiana. Int. Ver. Theor. Angew. Limnol. Verh. 16: WOLF, H.G., ANDG. A. CARVALHO Resting eggs of lake Daphnia. 2. In situ observations on the hatching of eggs and their contribution to population and community structure. Freshwater Biol. 22: 47 l AND M. A. MORT Inter-specific hybidization underlies phenotypic variability in Daphnia populations. Oecologia 68: WOLTERECK, R Races, associations and stratification of pelagic daphnids in some lakes of Wisconsin and other regions of the United States and Canada. Wis. Acad. Sci. 27: Submitted: 27 December 1990 Accepted: 5 September 1991 Revised: 21 October 1991 Limnol. Oceanogr.. 37(3), 1992, , by the American Society of Limnology and Oceanography, Inc. Cost of swimming by Daphnia during diel vertical migration Abstract-Daphnia hyalina was grown under simulated conditions of diurnal vertical migration in vertical flow chambers, and its life-history parameters were evaluated. The amplitudes of the migrations were 3 and 60 m d-l; food conditions were 0.1 and 1.0 mg C liter-. None of the investigated parameters (fecundity, individual growth rate, age of maturation) was significantly different between long- and short-distance migrating populations at high levels of food, Acknowledgments We thank Joanna Pijanowska, Barbara Taylor, and Winfried Lampert for helpful comments, and Nancy Weider for improving the English. whereas at low levels of food only the individual growth rate was higher in the short-distance migrating Daphnia. These results suggest that the energy spent to travel the vertical distance does not account for the metabolic costs of diurnal vertical migration. Several metabolic and demographic costs are associated with diel vertical migration (DVM) of zooplankton. Three components of these costs can be distinguished (see Lampert 1989): retarded metabolic rate due to migration through a vertical temperature